Goal congruency dominates reward value in accounting for behavioral and neural correlates of value-based decision-making

Abstract

When choosing between options, whether menu items or career paths, we can evaluate how rewarding each one will be, or how congruent it is with our current choice goal (e.g., to point out the best option or the worst one.). Past decision-making research interpreted findings through the former lens, but in these experiments the most rewarding option was always most congruent with the task goal (choosing the best option). It is therefore unclear to what extent expected reward vs. goal congruency can account for choice value findings. To deconfound these two variables, we performed three behavioral studies and an fMRI study in which the task goal varied between identifying the best vs. the worst option. Contrary to prevailing accounts, we find that goal congruency dominates choice behavior and neural activity. We separately identify dissociable signals of expected reward. Our findings call for a reinterpretation of previous research on value-based choice.

Fig. 1

Overall value effects on RT are driven by goal congruency rather than reward value. a After evaluating each item in isolation (left), participants saw sets of four options and (in separate blocks) were instructed to choose either the best or the worst option (right). The same example is shown for both blocks but each choice sets was only viewed once in a session. b Top: A reward-based account predicts that RTs should decrease with overall value of the set, irrespective of the choice goal. Bottom: A goal congruency account predicts that RTs should decrease with overall value in Choose-Best blocks but instead increase with overall value in Choose-Worst Blocks. c Both Study 1 (behavioral) and Study 2 (fMRI) find the task-specific RT reversal predicted by a goal congruency account (see also Supplementary Study 1, Supplementary Discussion). Shaded error bars show 95% confidence intervals. d Our empirical findings were captured by an LCA model that took goal values (rather than reward values) as inputs

Fig. 2

The valuation network tracks goal values and overall, but not relative reward value. a Valuation network mask. b Mixed-effects regression coefficients show that the valuation network ROI defined a priori based on ref. . tracks both overall and relative goal value, and also tracks overall (but not relative) reward value. Error bars show standard error of the mean. *p < 0.05, **p < 0.01

Fig. 3

Reward and goal value dissociate across the striatum’s dorsal-ventral axis. a Whole brain results for relative goal value (blue) and overall reward value (green), thresholded at voxelwise p < 0.001 and cluster-corrected p < 0.05. Despite being correlated with activity in our a priori value network ROI, overall goal value did not survive the whole-brain threshold used for these follow-up exploratory analyses. b To interrogate our findings across regions of striatum, we selected independent bilateral ROIs previously used as seeds for distinct resting-state networks: Dorsal Caudate (dark orange; [x, y, z] = ±12, 10, 8), Ventral Striatum, superior (orange; ±8, 10, 1), and Ventral Striatum, inferior (yellow; ±10, 11, −9). c Mixed-effects regression coefficients across these ROIs demonstrate a dorsal-ventral dissociation, with dorsal regions more sensitive to overall reward value and the ventralmost region more sensitive to relative goal value. Error bars show standard error of the mean. **p < 0.01, ***p < 0.001